From owner-freebsd-arch@FreeBSD.ORG Thu Aug 11 15:09:50 2005 Return-Path: X-Original-To: freebsd-arch@freebsd.org Delivered-To: freebsd-arch@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0208616A41F for ; Thu, 11 Aug 2005 15:09:50 +0000 (GMT) (envelope-from andre@freebsd.org) Received: from c00l3r.networx.ch (c00l3r.networx.ch [62.48.2.2]) by mx1.FreeBSD.org (Postfix) with ESMTP id D7F9443D45 for ; Thu, 11 Aug 2005 15:09:48 +0000 (GMT) (envelope-from andre@freebsd.org) Received: (qmail 26307 invoked from network); 11 Aug 2005 14:51:15 -0000 Received: from unknown (HELO freebsd.org) ([62.48.0.53]) (envelope-sender ) by c00l3r.networx.ch (qmail-ldap-1.03) with SMTP for ; 11 Aug 2005 14:51:15 -0000 Message-ID: <42FB6A3B.8A464F26@freebsd.org> Date: Thu, 11 Aug 2005 17:09:47 +0200 From: Andre Oppermann X-Mailer: Mozilla 4.8 [en] (Windows NT 5.0; U) X-Accept-Language: en MIME-Version: 1.0 To: John Baldwin References: <42F9ECF2.8080809@freebsd.org> <20050810.162006.48492066.imp@bsdimp.com> <200508110913.11867.jhb@FreeBSD.org> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: gnn@freebsd.org, freebsd-arch@freebsd.org Subject: Re: Special schedulers, one CPU only kernel, one only userland X-BeenThere: freebsd-arch@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussion related to FreeBSD architecture List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 11 Aug 2005 15:09:50 -0000 John Baldwin wrote: > > On Wednesday 10 August 2005 10:10 pm, gnn@freebsd.org wrote: > > But, the question, which is interesting to me for other reasons > > (namely putting in various real-time schedulers) was not answered. Do > > we currently have the requisite primitives to implement different > > schedulers or is there still a lot of jiggery/pokery required? > > Yes, there is some room for different schedulers, but the idea of using a > master/slave SMP system is not ust a different scheduler, but an entirely > different SMP architecture. That said, if you added one more sched_foo call > for synchronous kernel entry in trap() and syscall(), you probably could > provide a master/slave setup via a custom scheduler. It would be more > optimal to also tweak the low-level routing of interrupts to CPUs on > architectures that support it as well, though that would not be required. This is the answer I was looking for. Thanks. -- Andre